IBM

I haven't actually found a place for my 9114-275 yet, but it's definitely not gonna be in the middle of my dorm room stubbing my toe whenever it feels like it. I want to put it to some use, maybe even get some cash selling a shell account. It'll be running AIX 6.1 TL3SP1, with the usual things that might make a shell server useful to some people. I don't know how many people would be interested though. Would you?

Would it have gcc?

Well of course, seeing as I can't really see any use for a shell server other than to compile stuff on a weird and exotic platform. In fact I don't really get the draw of shell servers at all, that's why I want to set one up.

I'm also interested in finding a cheap copy of XLC.

ritchan wrote: I'm also interested in finding a cheap copy of XLC.

It doesn't exist. IBM isn't interested in hobbyists.

There's an academic program for professors, graduate research assistants, and high school teachers, but that's about it.

IBM did a trial copy of XLC which has a 90 day timeout. I just means I have to rebuild my AIX 5.3 server from scratch every 3 months!

Would you guys know enough about this to do it for yourselves? And shouldn't it be reinstall XLC, not reinstall AIX?
http://hardwarebug.org/2009/08/10/drm-the-big-blue-way/

Cool, I'll give that a go!

Okay, it's talking about ELF, my XLC, I am assuming, is XCOFF

It looks like he's talking about XLC 10.1, which is readily available from IBM at the moment. You're using XLC 9, eh? I don't know about ELF/COFF nor assembly, but what I do know is that I have had no luck getting the date out of 4a 0c 65 c8. IIRC, big endian means 8c 56 c0 a4, but when I told Windows calculator that it was a Dword, and typed it in both ways, I still got some kind of huge number.

So the function "_xlgetevalbeta()" returns a pointer to the data. The four bytes at offset 16 to that pointer are the "time_t" value for the time of installation. According to the "nm" map, the data is 20 bytes long, ( 0x14 )



So it looks like you could just compile a new library, rather than patch the old one.

(a) program to read 20 byte record from old library, to get all data
(b) fixup the current time
(c) write source file with new 20 byte data and function
(d) compile into library

ritchan wrote: 4a 0c 65 c8. IIRC, big endian means 8c 56 c0 a4

No, if it's big endian then 4A 0C 65 C8 means 4A0C65C8

If it's little endian then do the swap to C8650C4A

"ctime" will happily give you the string. As you can see my "4BD263E9" became "Fri Apr 23 22:22:17 2010".

[saya,shinichi]:~/source/Work $ CC time.cpp
[saya,shinichi]:~/source/Work $ ./a.out
Thu May 14 20:41:12 2009

After a lot of time spent at cplusplus.com I figured it out! Unfortunately it seems time_t won't accept hex values for a value, it only holds decimal. Well, that's what Windows calculator is for!

ritchan wrote: In fact I don't really get the draw of shell servers at all, that's why I want to set one up.

x86 "shell servers" (the kind you tend to see people paying money for online) are generally in datacenters and thus are used for anything that demands a stable connection but little resource and bandwidth, generally IRC bots/clients/servers or long-polling low-traffic network services like instant messenger bots.

Non-x86 "shell servers" are usually used to provide developers without ownership of the platform access to it to develop, for example Oregon State's Cell and POWER shell systems and OpenEmbedded.org's old ARM/XScale PDA cluster (which I think has been retired as Linux-based ARM devices propagated everywhere and demand for it waned).

Sometimes "shell servers" are used with distcc et. al. as a build box, but generally online services for this purpose are billed expressly as "build clusters," like the one Sourceforge gives access to and the Ubuntu package build system.

ritchan wrote: Would you guys know enough about this to do it for yourselves? And shouldn't it be reinstall XLC, not reinstall AIX?
http://hardwarebug.org/2009/08/10/drm-the-big-blue-way/

This is an awesome blog - thanks for that link. I wonder why the author remarks that because IBM's DRM is bad, their compiler must be bad as well though - it's a foolish generalization and discards most real-world facts about software licensing. Why waste time making something cracker-proof for a compiler that's so expensive that only straight-laced corporate users are going to purchase it anyway? And how does this naive licensing approach reflect negatively on the merit of compiler developers?

ritchan wrote: Unfortunately it seems time_t won't accept hex values for a value, it only holds decimal.

It holds neither decimal or hex, it holds a binary integer.

try





At least I know what project I'll be testing around July 23rd.

bri3d wrote:

ritchan wrote: Would you guys know enough about this to do it for yourselves? And shouldn't it be reinstall XLC, not reinstall AIX?
http://hardwarebug.org/2009/08/10/drm-the-big-blue-way/

This is an awesome blog - thanks for that link. I wonder why the author remarks that because IBM's DRM is bad, their compiler must be bad as well though - it's a foolish generalization and discards most real-world facts about software licensing. Why waste time making something cracker-proof for a compiler that's so expensive that only straight-laced corporate users are going to purchase it anyway? And how does this naive licensing approach reflect negatively on the merit of compiler developers?

IBM's problem is that they assume there are still honest people in this day and age. (har har)

That post read something like: "Wow! that house was easy to break into, Ionly had to cut two deadbolts. If these people don't bother putting a better lock on their door then I get to break in and do as I wish to their sleeping bodies."

It's all fun and games until you get an assfull of buckshot. I don't even want to think about how big the IBM legal department is... pirating software is one thing, posting a how-to on the internet is another.

You'll have an uphill battle making me believe that they care. But you're right - posting a howto is quite different from posting a crack. Which means it's even harder to find a case against him, don't you think?

sybrfreq wrote: That post read something like: "Wow! that house was easy to break into, Ionly had to cut two deadbolts. If these people don't bother putting a better lock on their door then I get to break in and do as I wish to their sleeping bodies."

It's all fun and games until you get an assfull of buckshot. I don't even want to think about how big the IBM legal department is... pirating software is one thing, posting a how-to on the internet is another.

I doubt he'll see any legal action. His blog post was pretentious and stupid, but the trial v. full version restrictions at that level are really just to allow engineers to evaluate and then remind them to ask for a purchase order - the piracy/loss risk is pretty low. Corporations like circumventing license restrictions when they can't tell they're doing something wrong (sharing logins to online databases, etc.) but as soon as they're clearly breaking things (i.e. cracking) someone in the process tends to straighten up (be it engineers or management). And it's been proven time and time again that IBM couldn't care less about hobbyists regardless of what they're doing.

I just think it's dumb to criticize compiler developers for writing simplistic/poor DRM when it's a low-priority, symbolic piece of the software, while simultaneously bragging about breaking it when it was dead-obvious how to do so. It's a silly stance I wouldn't have expected considering the intelligent posts elsewhere on that blog.

Right, I've never heard about that website before but "the license was easy to crack and therefore the software is crap" means I won't bother with it again.

sybrfreq wrote: Right, I've never heard about that website before but "the license was easy to crack and therefore the software is crap" means I won't bother with it again.

Huh? Do you realize you're doing the same thing as he did?

Anyway, I guess two pages of posts means it should be worth the hassle to set up an AIX 6.1 shell server. I think I'll try the post's method to get XLC 10.1 or 11.1 on it.